Kimi Code API 接入点不支持请求体压缩，长上下文场景下 TTFT 延迟显著

What feature would you like to see?

在使用 Kimi Code API（如 https://api.kimi.com/coding）时，我发现当前接入点不支持请求体（request body）的 gzip 或 deflate 压缩。在长上下文（例如 >50KB 的 prompt）场景下，未压缩的请求体会导致：- 网络传输时间增加- 首 Token 延迟（TTFT, Time To First Token）显著上升- 用户体验下降，尤其在高延迟网络环境中建议：在 API 层面支持 Content-Encoding: gzip 请求头，允许客户端发送压缩后的 JSON 请求体，以优化长上下文推理性能。此功能在 OpenAI、Anthropic 等主流 LLM API 中已广泛支持，对开发者体验和性能至关重要。

Additional information

No response

Start by tracing how requests are sent to the Kimi Code API endpoint at https://api.kimi.com/coding and determine whether this repository owns the endpoint or only the client request path. No source file or test is named; the work is complete when the supported compression behavior for large JSON request bodies is implemented or clearly documented and verified for long-context requests.